Getting to Know Your AI Agents: From Setup to Effective Usage

In our TeamPal user guide, we covered the basics of setting up your workspace and getting started with AI-powered agents. 

Now, it's time to take a deeper dive into the AI agent handbook, your go-to resource for mastering the creation, customization, and effective use of TeamPal agents. 

Whether you're selecting from a library of pre-designed templates or crafting an agent entirely from scratch, this guide is here to help you unlock their full potential and streamline your operations like never before.

Where Can You Get Your Agents?

Agent Library

To kick things off, head to the Agent Library from the main menu—it’s like the secret menu for AI aficionados. Here, you can browse over 80 templates designed for various departmental needs.

To add a new agent to your workspace, click on the agent profile you'd like to customize. Just a few clicks, plus credits, and voilà! The agent is yours!

Create a New Agent

If your needs are as unique as an Operations team’s coffee preferences (and we know how particular those can be), consider creating a new agent from scratch. 

Click Create agent on the homepage and tailor your agent’s role and tasks.

Customize the agent’s settings to match your specific needs:

After clicking Create new, your new agent will show up in the designated folder.

Set Up Your Agent

Now, you can customize your agent further to your needs. Let’s try giving Ava some makeover!

Click on the Gear icon to go to Agent configuration. 

Once you’re there, you’ll see five tabs: Instruction, Interface, Tasks, Super Power, and Publish. Let’s go over each tab.

1. Instruction

This tab includes the basic information of Ava, such as name, gender, position, job description, AI model, etc. 

Anytime you edit something in this tab, you have two saving options: Save & reset agent, which costs 30 credits and auto-generates all agent information based on your changes, and Save, which only saves your changes and keeps the other information the same.

2. Interface 

This tab is where you can set up the agent’s initial messages. Be creative and give Ava a humorous welcome message, if you’d like! ^^

3. Tasks

This tab allows you to add prompt templates for the tasks the agent might need to perform regularly, for easy access. 

The list of 20 tasks of each agent template you’ll see at first is automatically picked out from our 800+ task pool by our AI prompting technology. But if you’re creating a new agent from scratch, you’ll have to manually configure the task list from scratch. 

To further customize the list, you can either check out our task library… 

Or choose from your custom tasks in the Workspace task tab. 

Click Create to create new ones. You can preview it before saving!

Then, click Save and you’re good to go.

4. Super Power

Similar to the button you see in the agent chat screen, this tab is where you equip your agent with special skills, knowledge, and brand voice.

5. Publish

This tab is for your publishing needs. Publish your agents to share among the teams or use them as chatbots for customer support.

If you've encountered a problem with the AI agents, check out this article. If your problem hasn't been mentioned, leave us an email at [email protected]!

How to Use Your Agents Effectively

Once you’ve selected or created your AI agents, it’s crucial to know how to use them effectively to maximize their potential. Here are some key strategies to ensure you get the best results from your TeamPal agents:

Setting Up Your Agent Carefully

As mentioned earlier, each agent comes with specific instructions tailored to its role. However, for optimal performance, you’ll need to review and fine-tune certain elements:

• Agent description: Start by checking and revising your agent’s description. This is automatically generated by TeamPal based on your job brief, so ensure it aligns with your needs. It serves as the foundation for the agent’s functionality and helps clearly define its purpose and role within your team.

• System prompts: The instructions for each agent are generated based on the agent’s description and brand information. Make sure to review and refine these system prompts to ensure they perfectly align with your goals.

For example, if you want Zoe—an agent you’ve created—to focus on lead generation, include this in the system prompt and regenerate it for enhanced alignment.

• Agent model selection: Carefully select the AI model that best suits your agent’s tasks. Different models excel in different areas, so choosing the right one is essential to achieve your desired results. For details on selecting the most suitable models, check out this article.

By taking the time to set up your agents carefully and customizing these features, you ensure they are equipped to handle tasks effectively, enabling a seamless workflow and maximizing their potential within your organization.

Activating Your Agent’s Super Powers

To ensure your agent performs at its highest potential, it’s important to activate its relevant "Super Powers" before making any requests. These Super Powers include Skills, Knowledge, and Brand voice, which we have described to you in the previous section.

• Skills: Assign specialized abilities to your agent, such as conducting web searches, retrieving data from YouTube, or generating AI-powered images. These skills empower your agent to perform a variety of tasks efficiently and effectively.

• Knowledge: Equipping your agents with targeted knowledge ensures your agent has access to your business-specific data, allowing them to deliver precise and relevant answers. When making a request, remember to refer to that information in your prompts. For instance, “Based on our company’s Q2 sales data, what trends do you notice?” 

• Brand voice: Define and assign a unique brand voice to your agent that aligns with your company’s values and communication style. Your agent will generate responses consistent with your brand’s tone, ensuring a cohesive user experience.

Taking the time to activate relevant Super Powers ensures your agents are fully equipped for the tasks they need to handle, boosting their performance and effectiveness within your operations.

Optimizing Your Interaction with AI Agents

To maximize the potential of your TeamPal agent, how you communicate with it is as important as the setup process. Thoughtful interaction ensures precise, relevant, and actionable outputs. Here are several strategies to help you engage with your agent effectively:

Use Clear and Specific Prompts

The quality of the output your agent provides depends heavily on the clarity of your input. Always ensure your instructions are specific, detailed, and directly address your needs.

So, add necessary background information so the agent fully understands your request.

For example, instead of asking, “How can I improve my marketing?” try to ask, “What are three data-driven strategies to improve email open rates for a B2B SaaS company?”

Break Down Complex Tasks

Handling intricate requests can overwhelm even the most advanced AI. Simplify your tasks by breaking them into smaller, manageable steps for more accurate results.

For example, instead of asking for a complete product launch strategy in one go, break it into stages with prompts like:

• “What are the key components of a successful product launch strategy?”

• “How should we structure our pre-launch marketing campaign?”

• “What metrics should we track during the first month after launch?”

Utilize Follow-up Questions

Sometimes, your agent's initial response may not fully capture what you envisioned. Follow-up questions can enhance clarity and depth.

Use phrases like “Can you elaborate on…” or “Please provide more details about…” to get more comprehensive information.

For example, if you start with "What are the benefits of implementing a customer loyalty program?", follow up with “Can you provide specific examples of successful loyalty programs in the retail industry?”

Experiment with Different Prompting Techniques

Tailor your interaction style to suit your needs by experimenting with a variety of prompting techniques. Let’s try interacting with Ash, our marketing manager:

• Chain-of-thought: Guide the agent through a step-by-step reasoning process.

Sample prompt: 

"Let’s approach this step-by-step to develop a marketing strategy:

1. What are the current trends in digital marketing in 2024 relevant to SaaS companies?

2. How do these trends apply to our B2B target audience in the technology sector?

3. Based on this, suggest three key strategies we should focus on for our next quarter."

• Few-shot learning: Provide a few examples of the type of response you’re looking for.

Sample prompt: 

"I need help drafting LinkedIn posts for our upcoming webinar. Here are two examples of the tone and style I’m looking for:

Example 1:

🚀 Join us as we dive into cutting-edge SaaS trends of 2023!

Explore the future of innovation with industry experts in our free webinar. Register now!

Example 2:

📅 Mark your calendars!

This Wednesday, we’re unveiling strategies to optimize your business operations with AI-powered solutions. Don’t miss out—sign up today.

Now, could you draft two LinkedIn posts announcing our “AI for SaaS Success Webinar”? Make them engaging and on-brand."

• Hypothetical scenarios: Present a "what-if" scenario to encourage creative and actionable responses based on potential situations your business might encounter.

Sample prompt: 

“Imagine this scenario: We are planning to launch a new feature for our SaaS platform focused on AI automation. However, a competitor has already announced a similar feature just days before our launch.

What steps should we take to adjust our marketing strategy to differentiate our product and regain the attention of our target audience?”

Refine and Iterate

Don’t settle for unsatisfactory outputs—prompt refinement ensures ongoing improvement.

• Rephrase prompts: Provide additional details or reframe your request to guide the agent better.

• Learn from success: Observe which approaches yield the best responses and use them as templates for future prompts.

For instance, if your initial prompt is “Give me ideas for team-building activities.”, its refined version can be “Suggest 5 virtual team-building activities suitable for a remote marketing team of 15 people, with a focus on improving communication and creativity.”

Combine Human Insight with AI Capabilities

AI agents are powerful tools, but they work best when combined with human expertise. Treat them as collaborators rather than replacements.

• Critical evaluation: Analyze the agent’s outputs and refine them with your insights.

• Collaboration: Use the agent’s outputs as a foundation and build upon them with your knowledge to achieve better outcomes.

Keeping Your Agent Up-to-Date

To ensure optimal performance, it’s essential to regularly maintain and update your agent. AI agents thrive on current, relevant information and the latest capabilities.

• Update the knowledge base: Periodically review your agent’s knowledge base. Ensure all data, documents, and resources are accurate and reflect the latest information relevant to your business and objectives.

• Leverage platform updates: Stay informed about changes and updates within the TeamPal platform. These updates may introduce new features or enhancements that can further improve your agent’s functionality and effectiveness.

Upholding Ethical Standards

When using AI agents, maintaining ethical practices plays a crucial role in fostering trust, compliance, and responsible use.

• Follow organizational guidelines: Ensure your agents are used in alignment with your organization’s ethical policies and data protection standards. This includes adhering to industry regulations and internal codes of conduct.

• Protect sensitive information: Be cautious when providing your agents with private or confidential data. Safeguard personal and proprietary information and ensure ethical considerations are upheld to maintain security and trustworthiness in their usage.

These tips will help you harness the full potential of your TeamPal agents, leading to more efficient operations, better decision-making, and enhanced productivity across your organization. 

Ready to try creating and optimizing your AI workforce for yourself? Sign up for TeamPal today!